From 1976 to 1988, 496 patients with rectal adenocarcinoma have been treated at Gustave-Roussy Institute. There were 258 men (52%) and 238 women (48) with a median age of 61 years. Sixty pts (12%) had local treatment (contact-therapy or electro-resection). Thirty six pts (8.25%) had a simple exploration with colostomy. Four hundred pts had a resection: 202 abdomino-perineal resection, 7 perineal resection, 167 anterior resection, and 24 Hartmann's technique. Post-operative mortality was 1% (4 pts). Among these 400 pts, 208 had no complementary treatment, 134 had pre +/- post-operative radiotherapy and 58 had post-operative radiotherapy. Sixty one pts had palliative resection. The actuarial survival of the 400 pts at 3, 5 and 10 years are respectively 65%, 51% and 37.5%. The number of lost to follow-up patients was 11 (2.5%) at 5 years and 24 (5.5%) at 10 years. A retrospective uni and multifactorial analysis of the clinical, biological and histopathological data of the 400 pts was done, 18 factors were studied. Our judgement criterion was 5 year survival. The uni-factorial analysis showed 7 variables which had great influence on survival: age > 60 (p = 0.001), signs of severe illness (p < 0.0001), curative or palliative criterion of the surgery (p = 0.0001), depth of invasion (p = 0.0001), lymph node invasion (p = 0.0001), neural invasion (p = 0.0001) and positive emboli (p = 0.0001).(ABSTRACT TRUNCATED AT 250 WORDS)